Cosmology Using Photometric Samples of Type Ia Supernovae: The First Joint Photometric Light Curve Analysis

Abstract

Over the last twenty five years, type Ia supernova (SNIa) have been a crucial cosmological probe, responsible for the discovery of the accelerating expansion of the universe. But with the dawn of the next generation of telescopes and SNIa samples, new methods and techniques are needed to increase precision and handle ever-smaller systematics.The Dark Energy Survey (DES) Supernova program has recently completed, and will be the first program whose success relies on the analysis of a sample in which we do not know the typing of the supernova. These `photometric samples' are a significant switch from spectroscopic samples, in which every SNIa used in the analysis is spectroscopically identified. The motivation for this switch is the opportunity to have $\sim 10$ times larger samples than what would be feasible with spectroscopic identification; however, these analyses are replete with new and untested sources of systematic uncertainty. One less obvious, but equally significant, challenge is understanding the selection effects of this new survey strategy. This thesis serves to demonstrate the efficacy of these photometric samples. Simultaneously, as our statistical constraints improve, a higher burden is placed on other systematic uncertainties, like a better understanding of the environment that SNe occur in. Dust attenuates and reddens SNIa light curves, obscuring the true properties and astrophysical origins of the SNIa explosion. Here I combine the results of my four first-author papers in non-chronological order, laying out the steps I took to perform a cosmology analysis from start to finish. In Figure \ref{fig:roadmap}, I provide a general overview and outline of the course of this thesis, and how my papers aid in answering these questions. My first question is \textit{how do we observe the universe?} I go over the surveys and experiments that have been used to peer back billions of years into cosmic history. I also discuss how our imperfect instruments and selection effects all impact our observations, alongside how non-cosmological effects make these observations even more challenging, and how we try to mitigate those issues. To aid our methods of observing the universe, I ask - \textit{how do we simulate the universe?} These simulations are able to create extremely realistic, catalogue level samples that can be tuned to mirror our data in every measurable metric. This question leads to the first published work in this thesis: \textit{The Pantheon+ Analysis: Forward-Modeling the Dust and Intrinsic Colour Distributions of Type Ia Supernovae, and Quantifying their Impact on Cosmological Inferences}. The specific characteristics and modeling of dust distributions are drawn from the data using a multi-dimensional Markov Chain Monte Carlo method to infer and separate the intrinsic SNIa properties from those caused by external dust effects, informing us as to \textit{how our measurements are biased}.Given our improved simulations, I ask \textit{how do we fix these biases?}. In the next paper, \textit{Improved Treatment of Host-Galaxy Correlations in Cosmological Analyses With Type Ia Supernovae}, I introduce a method to fix these biases using simulations by providing the first phenomenological model of correlating SNIa properties with that of their host galaxy. This framework is used as the basis for a novel set of bias corrections that are able to account for realistic correlations between SNIa properties, as well as separately introducing the first bias corrections methodology to correct dust models of SNIa scatter. In the process of building up to cosmology measurements, I ask \textit{what biases are unique to photometric surveys?}. In \textit{Assessment of the Systematic Uncertainties in the Cosmological Analysis of the SDSS Supernovae Photometric Sample}, I investigate a collection of potential biases and uncertainties that are unique to photometric samples: assessing the impact of mis-associating the host galaxy, modeling of non-Ia contamination, and changing the modeled efficiency of detecting the host galaxies.My last paper puts all of these papers and methods together to answer the question: \textit{What cosmological result do we find?} I create the first-ever joint analysis of two photometric SNIa samples, making the largest SNIa analysis to-date. I test for consistency between the samples with the first comparison of statistically-independent SNIa samples, and show that the results of this joint photometric analysis are competitive with the best spectroscopic SNIa analysis.